How to judge if my work is meaningful in cryptography? [closed]

Suppose I have generated a new cryptographical result - for example a new cryptographical primitive, or a cryptanalytic attack. How can I find out whether this result is meaningful (significant)?

If the work has never been done by previous people, can I say it is meaningful?

-

closed as too broad by D.W., AFS, fgrieu, Gilles, archieJan 12 at 5:06

There are either too many possible answers, or good answers would be too long for this format. Please add details to narrow the answer set or to isolate an issue that can be answered in a few paragraphs.If this question can be reworded to fit the rules in the help center, please edit the question.

You might want to clarify what precisely you mean. Do you mean that, if you've done a trial decryption, how do you check if the result is meaningful? Or, do you mean that, if you've generated a cryptographical result (either a new cryptographical primitive, or a cryptanalytic attack), whether this result is meaningful (significant)? –  poncho Jan 6 at 4:09
I mean the latter that you present –  Alex Jan 6 at 7:26
Write it up and submit it to a suitable conference? –  DrLecter Jan 6 at 8:57
I still don't understand what the author means by "meaningful". Do you mean "correct"? "important"? "novel"? "useful"? "valuable"? "worth money"? something else? The question needs a lot more clarification. P.S. I'm not sure what is meant by "Can I say...?" You can say anything you want; it doesn't mean you'll be right, but you can say it. –  D.W. Jan 9 at 0:45
@D.W. : if you get approved by most of people in this field, then, your work will be meaningful, Isn't it? –  Alex Jan 10 at 14:01

With this amount of information it is hard to advice. Previously when somebody asked about new algorithm he had produced, he was answered: Answer to Where could I submit my algorithm?. No matter what kind of cryptographic work, generally large part of that answer applies.

If we knew anything about the work (like on which existing algorithms or problems you worked on), it could be possible to say if it can be unique work. Most likely people here only can judge that your work is:

• NOT relevant
• maybe relevant

They most likely cannot confirm the result to be unique and groundbreaking. In most cases, however, results appear to be not so unique.

If the work has never been done by previous people, can I say it is meaningful?

Most things people do have little significance in the big picture. Therefore, just being not done before does not mean it is meaningful. It is also impact and applicability of the result. Examples of results which could be important:

• A calculated SHA-1 collision.
• Faster than $2^{126.1}$ key recovery attack for AES.

Coming up with novel idea in cryptography is hard

The field of cryptography is very well thought of. Usually when confronted with a problem and I try to come up with a solution for it, I either:

• Think of a good practical solution
• Think of a broken solution
• Notice I cannot come up with a solution

In case of the first choice, I'm always able to find a pre-existing paper describing the solution. To come up with novel and unique solutions, it is really important to read and understand all of the relevant field, including the very recent cryptographic results, to come up with novel ideas.

Given that the algorithms used in practice are usually tens of years old, this may be counter-intuitive.

-
then what about the work I'm going to do ? –  Alex Jan 6 at 13:50

As with any research job, I can see two main ways that might allow you to find your work meaningful:

• Experts like it
• Company/state secrets
• The public uses it

The "correct" answer is that if you have come up with a great new result, you should try and get it published in one of the major conferences or similar event.

Alternatively, if you already work for a company with secret data (which could be a government or simply a big firm that mis-understands Crypto research) you might be happy with presenting them with your work and them using it. Take for example differential cryptanalysis. Whoever first discovered it for the NSA has certainly produced meaningful work, but the external credit goes to Biham-Shamir.

I put the public uses it as another point, but I don't think its a good thing. Some products (and thus their inventors work) are very important, but contain very bad crypto (eg MiFare Card ). In fact, there are lots and lots of such products, often from private companies that want to say "using our new specially developed crypto algorithm". The number of people that depending on such crypto demands they be accepted as meaningful, but that doesn't make them good!

-

The problem with new cryptographic algorithms is the amount of time and effort it takes to convince people of its security. There are many factors working against new development:

• Existing ciphers have been subjected to intense scrutiny, and yet are barely trusted.
• Changes to existing ciphers or protocols are viewed with suspicion. Is there a back door or "escrow" system built in? How can you prove there isn't?
• Many (most?) new ciphers, protocols, or security systems that have come along are broken sometime after an expensive hardware investment and deployment. See stories on WEP, HDCP, GSM, MiFARE, DVD CSS, SKIPJACK, LEAF, Dual_EC_DRBG, MD4, MD5, NTLM, and the list goes on.
• Existing primitives are already well served by adequate algorithms: secret key, public key, hashing, etc. Adding more algorithms does not automatically add benefit.

For your work to be recognized, you would need to come up with something truly outstanding that fills a recognized gap: perhaps something like nondeterministic random number generation (which would honestly be viewed by most people as a 'perpetual motion machine'.) But you can't swing a keyboard around here without hitting at least two people saying "I've got a new One Time Pad generator!" who don't understand that they have (poorly) reinvented a stream cipher. Don't be one of those people - they're not remembered, they're mocked.

You will likely have better luck making a name for yourself if you come up with something demonstrable, such as a break in an existing routine. To join the big leagues, you'd likely have to develop a novel class of cryptanalysis (such as Differential Analysis.)

-